50/50 my butt

Unknown

walkyourpath wrote:

I think these statistics are accurate. My butt is exactly 50/50. Half on the left and the other half on the right.

Can't speak for other players, they might be on the rear-end of the curve. I suppose with out the developers' insight on the backside, all our speculation is up in the derri-air.

wonderfully cheeky comment!

Unknown

Spencer75 wrote:

I have a strong background in statistics and I think you guys might be dismissing the OP a little too quickly.

He's received 34 awards. Leave out the 2 missions that only reward one prize ( the high iso prize)

18 other missions with the first one guaranteed.

So he's basically received 16 awards where he could have gotten a 20 iso reward instead.

Now he doesn't tell us how many times exactly he grinded these missions, but how many times before you would take his claims more seriously?

I don't know if we should dismiss these claims so quickly. How about collecting actual data?

I also feel I'm below 50% personally, but I know I've also gone on long successful streaks, and I know it's human nature to remember the bad streaks more. As Blue Shoes said, of all the people playing all the rounds, there are enough attempts that we expect *someone* to hit these awful streaks, and there's a selection effect that these people are more likely to take to the forums and vent than people are to post about going 20 straight times without a 20 ISO payout.

So, I guess I'm saying I don't really doubt him, and until I see some hard data my prior assumption is it's about equally likely he's just unlucky or the true rate is not 50/50 either due to coding bug or malice.

I think it would take about 10 people to report honestly and thoroughly on the next PVE tournament to actually get enough data to answer this question with reasonable certainty. Someone setup a spreadsheet to track what's pulled out of tokens. It would be cool if someone sets up another spreadsheet to track this. I laid out the (quite simple) math above, but I'm too lazy to actually setup the spreadsheet. Just count attempts and successes, compute the variance from the attempts, and keep a running tally of the observed success rate and how many standard deviations it is from a 50/50 success rate.

For extra credit, see if you can figure out how to deduce the most likely true success rate from the observed data. That would be a fun little introduction to Bayesian inference

Unknown

There is also the possibility that the awards generator pulls one of four prize potentials - and then takes the 50/50 (or vice versa - works the same the other way)

if there is a 50/50 you win a prize, (and the first prize is ALWAYS awarded), the AI / RNG pulls one prize from the four potentials - including the one already won.

declining percentages to clear the prize pool would be - three prizes - 0.375, two prizes 0.25, and the elusive last prize 0.125 chance (all non cumulative). Theoretically, that yields a statistical average of ~16 plays to get all four prizes. I have no hard empirical data, but that number "feels" right.

However, as Mark Twain once said:
"There are three types of lies - lies, damn lies, and statistics"

them's my thoughts

Unknown

I've been getting 20 iso all morning. After I get that first reward, I now expect a river of 20 ISO.

Unknown

Vinny: The game used to just award equal probability to all prizes, and if it picked one you've already won, it gives 20 ISO instead.

Some time back, it was changed so that after the first prize, it would give one of the remaining prizes 50% of the time, and 20 ISO the other 50% of the time. In other words, their claimed method is coin flip for new prize or 20 ISO, if new prize then choose any remaining prize with equal probability. The order of coin flip/choose a prize doesn't matter, but whether you choose from all possible prizes or only unrewarded prizes does make a difference.

This was supposed to improve the chances of getting a prize. However, it's pretty obvious that it actually decreases the odds of getting the 2nd prize (75% --> 50%), doesn't change the odds of getting the 3rd prize, but bumps the odds of the last prize from 25% --> 50%. So, I believe it's a modest improvement to the avg. number of tries to get all prizes, but not a huge bump (I haven't done that math).

Halfway through the 2nd sub event of the current PVE, I decided to keep a tally of all my attempts. Right now I'm batting 17/50, so about 33%. Obviously I'm well below pace if it's really 50/50, but I'd also be well ahead of the curve if your scenario were the case. It's still small sample size theatre at this point, so I'm not drawing conclusions yet. But, once I get a couple hundred attempts, I'll do another post with some findings.

Unknown

RNG definitely stuck today. 5 matches, all 20 ISO hooray!

Unknown

John Courage wrote:

Halfway through the 2nd sub event of the current PVE, I decided to keep a tally of all my attempts. Right now I'm batting 17/50, so about 33%. Obviously I'm well below pace if it's really 50/50, but I'd also be well ahead of the curve if your scenario were the case. It's still small sample size theatre at this point, so I'm not drawing conclusions yet. But, once I get a couple hundred attempts, I'll do another post with some findings.

I will keep track as well on the next set of subs - nothing like empirical data to back a theory. heck, I'll even post my findings if I am >* cough*< wr... - less correct than I could be

for more data, I will track # of attempts to 2nd award, then to 3rd, etc, That will still generate meaningful data without grinding all missions to completion. I am still thinking this is a game for fun, not so much a job...

Spencer75

Vinny J wrote:

John Courage wrote:

Halfway through the 2nd sub event of the current PVE, I decided to keep a tally of all my attempts. Right now I'm batting 17/50, so about 33%. Obviously I'm well below pace if it's really 50/50, but I'd also be well ahead of the curve if your scenario were the case. It's still small sample size theatre at this point, so I'm not drawing conclusions yet. But, once I get a couple hundred attempts, I'll do another post with some findings.

I will keep track as well on the next set of subs - nothing like empirical data to back a theory. heck, I'll even post my findings if I am >* cough*< wr... - less correct than I could be

for more data, I will track # of attempts to 2nd award, then to 3rd, etc, That will still generate meaningful data without grinding all missions to completion. I am still thinking this is a game for fun, not so much a job...

This

If we could find reliable people (like Vinny J and John courage) to record their results we can start to get at the truth here, rather than dismissing things out of hand.

I think it's worth investigating.

Vinny and John Courage clearly have a mind for this, but you obviously can't just take any form goer's input. People will be more inclined to report their bad results.

Spencer75

I have one reward for every battle and haven't attempted any of them a second time. I'm going to go in and start recording data from here on out.

Unknown

I can record all of my games from now on, as I am very curious about this.

For a while I even had a theory that time spent and/or damage received made a 20 more possible (since my highest is Daken, the games can end within a minute and there will be 0 damage on him - a 20 happened really often in such cases).

Unknown

It probably is 50/50. 50/50 means that on average it should take 6 tries of a mission after the first, guaranteed, reward to get all 3 rewards. I would rarely grind a mission that many times unless it's open for 2 days or more and keeps resetting in points (hello Simulator!) and wow, I almost never get all the rewards before a mission isn't worth the points-except in Simulator.

That said, it is definitely annoying when you get a streak like grinding a mission down until it is worth 1 point, still taking a couple more shots at it hoping for rewards, and still have it stuck at 1/4 rewards. Just random bad luck though.

I think they should increase the reward percentage, though. Asking people to do a mission on average 7 times to get all the rewards, with a reasonable chance of doing far more than that and still not getting them, sucks. Maybe if the consolation prize was more than a mere 20 iso (I can live with it if the 20 iso stays in for Prologue missions and/or for missions with Trivial difficulty).

Unknown

No matter what you record it doesn't change that your chances are 50/50 that you're going to get an insulting, discouraging 20 ISO.

Unknown

Anyone recall a case playing 4 times and get all 4 rewards?
I can't.

While getting 20 iso 4 or more times in a row is a pretty common experience.

It definitely is not 50/50. Originally I thought it gives an even chance to get any prize slot, and grants the 20 if you drew a taken one. But those insane streaks keep coming for 1/4 cases too.

It'll be interesting to see recorded data.

Unknown

On another note I remember IceIX stating that they didn't touch drop odds for boosts, but I haven't gotten a boost drop from a Prologue mission since R47 hit ages ago. And overall boost drops have decreased massively as well.

Spencer75

pasa_ wrote:

Anyone recall a case playing 4 times and get all 4 rewards?
I can't.

While getting 20 iso 4 or more times in a row is a pretty common experience.

It definitely is not 50/50. Originally I thought it gives an even chance to get any prize slot, and grants the 20 if you drew a taken one. But those insane streaks keep coming for 1/4 cases too.

It'll be interesting to see recorded data.

He's got a point here

If it really is 50/50 you would expect to sweep thru and go 4/4 12.5% of the time.

Unknown

You would have 12.5% chance of getting 4/4 because the first one is guaranteed.

I don't really pay attention to my rewards, but that'd mean you should have something with a check mark before hitting the 1 (or actually 1 mission before hitting the 1), and that doesn't sound like anything outrageous. I'm pretty sure I've seen this before.

On the current sub bracket I only made one pass so far, and here's the reward list for the missions I've done to 1 point (so I beat them 5 times total):

1/4
2/4
3/4
4/4 X 2

Statistically 3 should be the average (1st clear = guaranteed, next 4 clear = 50%). My average here is 2.8/4, so I see nothing wrong here.

Spencer75

Phantron wrote:

You would have 12.5% chance of getting 4/4 because the first one is guaranteed.

I don't really pay attention to my rewards, but that'd mean you should have something with a check mark before hitting the 1 (or actually 1 mission before hitting the 1), and that doesn't sound like anything outrageous. I'm pretty sure I've seen this before.

On the current sub bracket I only made one pass so far, and here's the reward list for the missions I've done to 1 point (so I beat them 5 times total):

1/4
2/4
3/4
4/4 X 2

Statistically 3 should be the average (1st clear = guaranteed, next 4 clear = 50%). My average here is 2.8/4, so I see nothing wrong here.

Outrageously small sample size alert

Unknown

Spencer75 wrote:

Phantron wrote:

You would have 12.5% chance of getting 4/4 because the first one is guaranteed.

I don't really pay attention to my rewards, but that'd mean you should have something with a check mark before hitting the 1 (or actually 1 mission before hitting the 1), and that doesn't sound like anything outrageous. I'm pretty sure I've seen this before.

On the current sub bracket I only made one pass so far, and here's the reward list for the missions I've done to 1 point (so I beat them 5 times total):

1/4
2/4
3/4
4/4 X 2

Statistically 3 should be the average (1st clear = guaranteed, next 4 clear = 50%). My average here is 2.8/4, so I see nothing wrong here.

Outrageously small sample size alert

He was just addressing a question, not claiming anything in general.

As to that question yeah, I've seen 4 straight non-20iso awards plenty of times. And 4 straight 20iso awards plenty of times as well. And everything in between plenty of times.

But I've never kept track and until one of us does none of this means anything. Until multiple people do over a long period of time (maybe I'll start) I have no reason to believe it's anything other than 50/50 based on my own play.

Unknown

pasa_ wrote:

Anyone recall a case playing 4 times and get all 4 rewards?
I can't.

While getting 20 iso 4 or more times in a row is a pretty common experience.

It definitely is not 50/50. Originally I thought it gives an even chance to get any prize slot, and grants the 20 if you drew a taken one. But those insane streaks keep coming for 1/4 cases too.

It'll be interesting to see recorded data.

This forum is so hardcore. I love it. Maybe a nifty spreadsheet is in order to keep a tally of everyones results? Like the standard token spreadsheet?

Unknown

My sample size of 25 matches disproves an outrageous claim because it's the guy making the outrageous claim that is supposed to provide the data, not me.

Let's say I have a random quarter, and you said you think the coin is rigged and not a fair coin, so I flip it 10 times and it landed head 6 times. That is good enough to disprove your theory because there's no reason for anyone to believe a random quarter is rigged and we didn't get any outrageous results from flipping it 10 times. If someone wants to show that the coin is indeed biased, they're free to run their own experiments to show it. The burden of proof is on them, not me.